这一篇写写复制错误处理相关的另两个参数slave_skip_errors、slave_exec_mode，基本环境参考《复制错误处理——sql_slave_skip_counter》
一、slave_skip_errors
1.1、slave_skip_errors官方解释
https://dev.mysql.com/doc/refman/5.7/en/replication-options-slave.html
Normally, replication stops when an error occurs on the slave, which gives you the opportunity to resolve the inconsistency in the data manually. This option causes the slave SQL thread to continue replication when a statement returns any of the errors listed in the option value.--slave-skip-errors=[err_code1,err_code2,...|all|ddl_exist_errors]
Do not use this option unless you fully understand why you are getting errors. If there are no bugs in your replication setup and client programs, and no bugs in MySQL itself, an error that stops replication should never occur. Indiscriminate use of this option results in slaves becoming hopelessly out of synchrony with the master, with you having no idea why this has occurred.
1.2、测试数据
slave_skip_errors是一个非Dynamic变量，在配置文件中添加
[mysqld] slave_skip_errors=1032,1062
主库创建一个事务表和一个非事务表，然后从库往各表写入id=1的记录
# 主库创建测试表 mydba@192.168.85.132,3306 [replcrash]> create table repl_innodb(id int primary key,name1 char(10),name2 char(10)) engine=innodb; mydba@192.168.85.132,3306 [replcrash]> create table repl_myisam(id int primary key,name1 char(10),name2 char(10)) engine=myisam; # 从库往测试表中添加数据，不记入binlog mydba@192.168.85.133,3306 [replcrash]> set sql_log_bin=0; mydba@192.168.85.133,3306 [replcrash]> insert into repl_innodb(id,name1,name2) values(1,'s1062-1','s1062-1'); mydba@192.168.85.133,3306 [replcrash]> insert into repl_myisam(id,name1,name2) values(1,'s1062-1','s1062-1'); mydba@192.168.85.133,3306 [replcrash]> set sql_log_bin=1;
1.3、一个事务中包含事务表和非事务表操作
这里不再单独对事务表和非事务表进行测试
# 主库往事务表、非事务表中添加数据 mydba@192.168.85.132,3306 [replcrash]> begin; mydba@192.168.85.132,3306 [replcrash]> insert into repl_innodb(id,name1,name2) values(1,'m1062-1','m1062-1'); mydba@192.168.85.132,3306 [replcrash]> insert into repl_innodb(id,name1,name2) values(2,'m1062-2','m1062-2'); mydba@192.168.85.132,3306 [replcrash]> insert into repl_myisam(id,name1,name2) values(1,'m1062-1','m1062-1'); mydba@192.168.85.132,3306 [replcrash]> insert into repl_myisam(id,name1,name2) values(2,'m1062-2','m1062-2'); mydba@192.168.85.132,3306 [replcrash]> commit; # 主库数据 mydba@192.168.85.132,3306 [replcrash]> select * from repl_innodb; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 1 | m1062-1 | m1062-1 | | 2 | m1062-2 | m1062-2 | +----+---------+---------+ mydba@192.168.85.132,3306 [replcrash]> select * from repl_myisam; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 1 | m1062-1 | m1062-1 | | 2 | m1062-2 | m1062-2 | +----+---------+---------+ # 从库数据 mydba@192.168.85.133,3306 [replcrash]> select * from repl_innodb; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 1 | s1062-1 | s1062-1 | | 2 | m1062-2 | m1062-2 | +----+---------+---------+ mydba@192.168.85.133,3306 [replcrash]> select * from repl_myisam; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 1 | s1062-1 | s1062-1 | | 2 | m1062-2 | m1062-2 | +----+---------+---------+
复制正常，从库repl_myisam表跳过id=1的记录，复制了id=2的记录；从库repl_innodb表跳过id=1的记录，复制了id=2的记录
语句操作过程：开启显式事务1，往repl_innodb表写入id=1、2两条记录-->开启事务2，往repl_myisam表写入id=1记录，提交事务2-->开启事务3，往repl_myisam表写入id=2记录，提交事务3-->提交显式事务1
当事务2提交后，从库上repl_myisam.id=1的Duplicate entry被skip；当事务3提交后，从库写入repl_myisam.id=2的记录；当事务1提交后，从库上repl_innodb.id=1的Duplicate entry被skip，从库写入repl_innodb.id=2的记录~
也就是说配置文件中的slave_skip_errors=1032,1062 它仅跳过出错的行，并不是跳过整个事务(sql_slave_skip_counter会跳过整个事务)
试想，如果主库上的显式事务中有update操作，然后在从库找不到对应行，它仅跳过出错的行，主从不一致的情况将继续，并且不会触发错误
[root@ZST1 logs]# mysqlbinlog -v --base64-output=decode-rows mysql-bin.000008 --start-position=1333 /*!50530 SET @@SESSION.PSEUDO_SLAVE_MODE=1*/; /*!50003 SET @OLD_COMPLETION_TYPE=@@COMPLETION_TYPE,COMPLETION_TYPE=0*/; DELIMITER /*!*/; # at 1333 #171201 21:08:18 server id 1323306 end_log_pos 1398 CRC32 0xd792f990 Anonymous_GTID last_committed=5 sequence_number=6 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 1398 #171201 21:08:18 server id 1323306 end_log_pos 1475 CRC32 0x1869ed89 Query thread_id=3 exec_time=0 error_code=0 SET TIMESTAMP=1512133698/*!*/; SET @@session.pseudo_thread_id=3/*!*/; SET @@session.foreign_key_checks=1, @@session.sql_auto_is_null=0, @@session.unique_checks=1, @@session.autocommit=1/*!*/; SET @@session.sql_mode=1436549152/*!*/; SET @@session.auto_increment_increment=1, @@session.auto_increment_offset=1/*!*/; /*!\C utf8 *//*!*/; SET @@session.character_set_client=33,@@session.collation_connection=33,@@session.collation_server=33/*!*/; SET @@session.lc_time_names=0/*!*/; SET @@session.collation_database=DEFAULT/*!*/; ==================== repl_myisam表写入id=1的记录Start ==================== BEGIN /*!*/; # at 1475 #171201 21:08:18 server id 1323306 end_log_pos 1540 CRC32 0x7f6a1e44 Table_map: `replcrash`.`repl_myisam` mapped to number 307 # at 1540 #171201 21:08:18 server id 1323306 end_log_pos 1596 CRC32 0xb6784f59 Write_rows: table id 307 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_myisam` ### SET ### @1=1 ### @2='m1062-1' ### @3='m1062-1' # at 1596 #171201 21:08:18 server id 1323306 end_log_pos 1674 CRC32 0xdebe509f Query thread_id=3 exec_time=0 error_code=0 SET TIMESTAMP=1512133698/*!*/; COMMIT /*!*/; # at 1674 ==================== repl_myisam表写入id=1的记录End ==================== #171201 21:08:40 server id 1323306 end_log_pos 1739 CRC32 0x09f01ffa Anonymous_GTID last_committed=6 sequence_number=7 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 1739 #171201 21:08:40 server id 1323306 end_log_pos 1816 CRC32 0x9b1cba09 Query thread_id=3 exec_time=0 error_code=0 SET TIMESTAMP=1512133720/*!*/; BEGIN /*!*/; # at 1816 #171201 21:08:40 server id 1323306 end_log_pos 1881 CRC32 0xf8e7ddd7 Table_map: `replcrash`.`repl_myisam` mapped to number 307 # at 1881 #171201 21:08:40 server id 1323306 end_log_pos 1937 CRC32 0xa2150d71 Write_rows: table id 307 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_myisam` ### SET ### @1=2 ### @2='m1062-2' ### @3='m1062-2' # at 1937 #171201 21:08:40 server id 1323306 end_log_pos 2015 CRC32 0xb007bae6 Query thread_id=3 exec_time=0 error_code=0 SET TIMESTAMP=1512133720/*!*/; COMMIT /*!*/; # at 2015 ==================== repl_myisam表写入id=2的记录End ==================== #171201 21:08:56 server id 1323306 end_log_pos 2080 CRC32 0x39e67db5 Anonymous_GTID last_committed=7 sequence_number=8 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 2080 #171201 21:07:59 server id 1323306 end_log_pos 2157 CRC32 0xb1ae59f2 Query thread_id=3 exec_time=0 error_code=0 SET TIMESTAMP=1512133679/*!*/; ==================== repl_innodb表写入id=1、2的记录，在一个事务中Start ==================== BEGIN /*!*/; # at 2157 #171201 21:07:59 server id 1323306 end_log_pos 2222 CRC32 0x09d40a4f Table_map: `replcrash`.`repl_innodb` mapped to number 306 # at 2222 #171201 21:07:59 server id 1323306 end_log_pos 2278 CRC32 0x834f2f78 Write_rows: table id 306 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_innodb` ### SET ### @1=1 ### @2='m1062-1' ### @3='m1062-1' # at 2278 #171201 21:08:07 server id 1323306 end_log_pos 2343 CRC32 0x5b7e244b Table_map: `replcrash`.`repl_innodb` mapped to number 306 # at 2343 #171201 21:08:07 server id 1323306 end_log_pos 2399 CRC32 0x965812b9 Write_rows: table id 306 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_innodb` ### SET ### @1=2 ### @2='m1062-2' ### @3='m1062-2' # at 2399 #171201 21:08:56 server id 1323306 end_log_pos 2430 CRC32 0xbddc1af8 Xid = 1076 COMMIT/*!*/; ==================== repl_innodb表写入id=1、2的记录，在一个事务中End ==================== SET @@SESSION.GTID_NEXT= 'AUTOMATIC' /* added by mysqlbinlog */ /*!*/; DELIMITER ; # End of log file /*!50003 SET COMPLETION_TYPE=@OLD_COMPLETION_TYPE*/; /*!50530 SET @@SESSION.PSEUDO_SLAVE_MODE=0*/; [root@ZST1 logs]#
注意：如果主库使用insert into repl_innodb(id,name1,name2) values(1,'m1062-1','m1062-1'),(2,'m1062-2','m1062-2');
复制正常，从库直接跳过整个事务，主上id=1、2记录不会插入到从库，这种写法id=1和id=2之间是没有position分隔(解析binlog可以看出)
二、slave_exec_mode
2.1、slave_exec_mode官方解释
https://dev.mysql.com/doc/refman/5.7/en/replication-options-slave.html
Controls how a slave thread resolves conflicts and errors during replication. IDEMPOTENT mode causes suppression of duplicate-key and no-key-found errors; STRICT means no such suppression takes place.SET GLOBAL slave_exec_mode = ['IDEMPOTENT'|'STRICT']
IDEMPOTENT mode is intended for use in multi-master replication, circular replication, and some other special replication scenarios for NDB Cluster Replication.
For storage engines other than NDB, IDEMPOTENT mode should be used only when you are absolutely sure that duplicate-key errors and key-not-found errors can safely be ignored. It is meant to be used in fail-over scenarios for NDB Cluster where multi-master replication or circular replication is employed, and is not recommended for use in other cases.
2.2、初始数据
注释配置文件中的slave_skip_errors，然后初始数据
# 主库数据 mydba@192.168.85.132,3306 [replcrash]> select * from repl_innodb; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 2 | m1032-2 | m1032-2 | | 3 | m1032-3 | m1032-3 | +----+---------+---------+ mydba@192.168.85.132,3306 [replcrash]> select * from repl_myisam; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 2 | m1032-2 | m1032-2 | | 3 | m1032-3 | m1032-3 | +----+---------+---------+ # 从库数据 mydba@192.168.85.133,3306 [replcrash]> select * from repl_innodb; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 1 | s1062-1 | s1062-1 | +----+---------+---------+ mydba@192.168.85.133,3306 [replcrash]> select * from repl_myisam; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 1 | s1062-1 | s1062-1 | +----+---------+---------+
这里只是为了模拟1062(insert遇到duplicate-key)、1032(delete/update遇到no-key-found)错误
2.3、一个事务中包含事务表和非事务表操作
这里不再单独对事务表和非事务表进行测试
# 主库开启事务，insert、update、delete mydba@192.168.85.132,3306 [replcrash]> begin; mydba@192.168.85.132,3306 [replcrash]> insert into repl_innodb(id,name1,name2) values(1,'m1062-1','m1062-1'); mydba@192.168.85.132,3306 [replcrash]> update repl_innodb set name1='m1032-2upd' where id = 2; mydba@192.168.85.132,3306 [replcrash]> delete from repl_innodb where id=3; mydba@192.168.85.132,3306 [replcrash]> insert into repl_innodb(id,name1,name2) values(4,'m1062-4','m1062-4'); mydba@192.168.85.132,3306 [replcrash]> insert into repl_myisam(id,name1,name2) values(1,'m1062-1','m1062-1'); mydba@192.168.85.132,3306 [replcrash]> update repl_myisam set name1='m1032-2upd' where id = 2; mydba@192.168.85.132,3306 [replcrash]> delete from repl_myisam where id=3; mydba@192.168.85.132,3306 [replcrash]> insert into repl_myisam(id,name1,name2) values(4,'m1062-4','m1062-4'); mydba@192.168.85.132,3306 [replcrash]> commit; # 主库数据 mydba@192.168.85.132,3306 [replcrash]> select * from repl_innodb; +----+------------+---------+ | id | name1 | name2 | +----+------------+---------+ | 1 | m1062-1 | m1062-1 | | 2 | m1032-2upd | m1032-2 | | 4 | m1062-4 | m1062-4 | +----+------------+---------+ mydba@192.168.85.132,3306 [replcrash]> select * from repl_myisam; +----+------------+---------+ | id | name1 | name2 | +----+------------+---------+ | 1 | m1062-1 | m1062-1 | | 2 | m1032-2upd | m1032-2 | | 4 | m1062-4 | m1062-4 | +----+------------+---------+
从库在应用insert into repl_myisam(id,name1,name2) values(1,'m1062-1','m1062-1')时，从库报错replcrash.repl_myisam; Duplicate entry '1'，SQL thread停止。前面对repl_innodb的操作还没有commit，所以从库此时数据没有任何变化。尝试设置从库设置slave_exec_mode参数
# 从库设置slave_exec_mode='IDEMPOTENT' mydba@192.168.85.133,3306 [(none)]> set global slave_exec_mode='IDEMPOTENT'; mydba@192.168.85.133,3306 [(none)]> start slave sql_thread; # 从库数据 mydba@192.168.85.133,3306 [replcrash]> select * from repl_innodb; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 1 | m1062-1 | m1062-1 | | 4 | m1062-4 | m1062-4 | +----+---------+---------+ mydba@192.168.85.133,3306 [replcrash]> select * from repl_myisam; +----+---------+---------+ | id | name1 | name2 | +----+---------+---------+ | 1 | m1062-1 | m1062-1 | | 4 | m1062-4 | m1062-4 | +----+---------+---------+
从库更新了id=1的记录，写入了id=4的记录，它是以什么顺序来执行的呢？我们先来查看主库上的binlog
[root@ZST1 logs]# mysqlbinlog -v --base64-output=decode-rows mysql-bin.000008 --start-position=2430 /*!50530 SET @@SESSION.PSEUDO_SLAVE_MODE=1*/; /*!50003 SET @OLD_COMPLETION_TYPE=@@COMPLETION_TYPE,COMPLETION_TYPE=0*/; DELIMITER /*!*/; # at 2430 #171201 23:01:19 server id 1323306 end_log_pos 2495 CRC32 0x99557a42 Anonymous_GTID last_committed=8 sequence_number=9 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 2495 #171201 23:01:19 server id 1323306 end_log_pos 2572 CRC32 0xdcd2d27f Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140479/*!*/; SET @@session.pseudo_thread_id=13/*!*/; SET @@session.foreign_key_checks=1, @@session.sql_auto_is_null=0, @@session.unique_checks=1, @@session.autocommit=1/*!*/; SET @@session.sql_mode=1436549152/*!*/; SET @@session.auto_increment_increment=1, @@session.auto_increment_offset=1/*!*/; /*!\C utf8 *//*!*/; SET @@session.character_set_client=33,@@session.collation_connection=33,@@session.collation_server=33/*!*/; SET @@session.lc_time_names=0/*!*/; SET @@session.collation_database=DEFAULT/*!*/; ==================== repl_myisam表写入id=1的记录Start ==================== BEGIN /*!*/; # at 2572 #171201 23:01:19 server id 1323306 end_log_pos 2637 CRC32 0x48b4dd17 Table_map: `replcrash`.`repl_myisam` mapped to number 309 # at 2637 #171201 23:01:19 server id 1323306 end_log_pos 2693 CRC32 0x8d519b60 Write_rows: table id 309 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_myisam` ### SET ### @1=1 ### @2='m1062-1' ### @3='m1062-1' # at 2693 #171201 23:01:19 server id 1323306 end_log_pos 2771 CRC32 0xe0be0f09 Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140479/*!*/; COMMIT /*!*/; # at 2771 ==================== repl_myisam表写入id=1的记录End ==================== #171201 23:01:36 server id 1323306 end_log_pos 2836 CRC32 0xfb8a1915 Anonymous_GTID last_committed=9 sequence_number=10 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 2836 #171201 23:01:36 server id 1323306 end_log_pos 2913 CRC32 0xf2abbe32 Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140496/*!*/; ==================== repl_myisam表更新id=2的记录Start ==================== BEGIN /*!*/; # at 2913 #171201 23:01:36 server id 1323306 end_log_pos 2978 CRC32 0x2b4285a4 Table_map: `replcrash`.`repl_myisam` mapped to number 309 # at 2978 #171201 23:01:36 server id 1323306 end_log_pos 3059 CRC32 0x6f94f7cb Update_rows: table id 309 flags: STMT_END_F ### UPDATE `replcrash`.`repl_myisam` ### WHERE ### @1=2 ### @2='m1032-2' ### @3='m1032-2' ### SET ### @1=2 ### @2='m1032-2upd' ### @3='m1032-2' # at 3059 #171201 23:01:36 server id 1323306 end_log_pos 3137 CRC32 0x2a8e5489 Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140496/*!*/; COMMIT /*!*/; # at 3137 ==================== repl_myisam表更新id=2的记录End ==================== #171201 23:01:45 server id 1323306 end_log_pos 3202 CRC32 0xc7ef1e5e Anonymous_GTID last_committed=10 sequence_number=11 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 3202 #171201 23:01:45 server id 1323306 end_log_pos 3279 CRC32 0xcf5e63ac Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140505/*!*/; ==================== repl_myisam表删除id=3的记录Start ==================== BEGIN /*!*/; # at 3279 #171201 23:01:45 server id 1323306 end_log_pos 3344 CRC32 0xb6020c12 Table_map: `replcrash`.`repl_myisam` mapped to number 309 # at 3344 #171201 23:01:45 server id 1323306 end_log_pos 3400 CRC32 0x5fc1ee98 Delete_rows: table id 309 flags: STMT_END_F ### DELETE FROM `replcrash`.`repl_myisam` ### WHERE ### @1=3 ### @2='m1032-3' ### @3='m1032-3' # at 3400 #171201 23:01:45 server id 1323306 end_log_pos 3478 CRC32 0x6aab62fd Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140505/*!*/; COMMIT /*!*/; # at 3478 ==================== repl_myisam表删除id=3的记录End ==================== #171201 23:01:52 server id 1323306 end_log_pos 3543 CRC32 0xed36610c Anonymous_GTID last_committed=11 sequence_number=12 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 3543 #171201 23:01:52 server id 1323306 end_log_pos 3620 CRC32 0x5a134f60 Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140512/*!*/; ==================== repl_myisam表写入id=4的记录Start ==================== BEGIN /*!*/; # at 3620 #171201 23:01:52 server id 1323306 end_log_pos 3685 CRC32 0x317d7f86 Table_map: `replcrash`.`repl_myisam` mapped to number 309 # at 3685 #171201 23:01:52 server id 1323306 end_log_pos 3741 CRC32 0x50f84020 Write_rows: table id 309 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_myisam` ### SET ### @1=4 ### @2='m1062-4' ### @3='m1062-4' # at 3741 #171201 23:01:52 server id 1323306 end_log_pos 3819 CRC32 0xdc200995 Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140512/*!*/; COMMIT /*!*/; # at 3819 ==================== repl_myisam表写入id=4的记录End ==================== #171201 23:01:58 server id 1323306 end_log_pos 3884 CRC32 0xaae63436 Anonymous_GTID last_committed=12 sequence_number=13 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 3884 #171201 23:00:04 server id 1323306 end_log_pos 3961 CRC32 0xb14f2893 Query thread_id=13 exec_time=0 error_code=0 SET TIMESTAMP=1512140404/*!*/; ==================== repl_innodb表操作id=1、2、3、4的记录，在一个事务中Start ==================== BEGIN /*!*/; # at 3961 #171201 23:00:04 server id 1323306 end_log_pos 4026 CRC32 0x62020112 Table_map: `replcrash`.`repl_innodb` mapped to number 308 # at 4026 #171201 23:00:04 server id 1323306 end_log_pos 4082 CRC32 0xa6c81e06 Write_rows: table id 308 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_innodb` ### SET ### @1=1 ### @2='m1062-1' ### @3='m1062-1' # at 4082 #171201 23:00:14 server id 1323306 end_log_pos 4147 CRC32 0x0a4abc3c Table_map: `replcrash`.`repl_innodb` mapped to number 308 # at 4147 #171201 23:00:14 server id 1323306 end_log_pos 4228 CRC32 0xdcd94aa2 Update_rows: table id 308 flags: STMT_END_F ### UPDATE `replcrash`.`repl_innodb` ### WHERE ### @1=2 ### @2='m1032-2' ### @3='m1032-2' ### SET ### @1=2 ### @2='m1032-2upd' ### @3='m1032-2' # at 4228 #171201 23:00:22 server id 1323306 end_log_pos 4293 CRC32 0xf5a4c9ba Table_map: `replcrash`.`repl_innodb` mapped to number 308 # at 4293 #171201 23:00:22 server id 1323306 end_log_pos 4349 CRC32 0xbd1864f7 Delete_rows: table id 308 flags: STMT_END_F ### DELETE FROM `replcrash`.`repl_innodb` ### WHERE ### @1=3 ### @2='m1032-3' ### @3='m1032-3' # at 4349 #171201 23:00:32 server id 1323306 end_log_pos 4414 CRC32 0x89bcb330 Table_map: `replcrash`.`repl_innodb` mapped to number 308 # at 4414 #171201 23:00:32 server id 1323306 end_log_pos 4470 CRC32 0x80d6ed1a Write_rows: table id 308 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_innodb` ### SET ### @1=4 ### @2='m1062-4' ### @3='m1062-4' # at 4470 #171201 23:01:58 server id 1323306 end_log_pos 4501 CRC32 0xdc2e8ab8 Xid = 1133 COMMIT/*!*/; ==================== repl_innodb表操作id=1、2、3、4的记录，在一个事务中Start ==================== SET @@SESSION.GTID_NEXT= 'AUTOMATIC' /* added by mysqlbinlog */ /*!*/; DELIMITER ; # End of log file /*!50003 SET COMPLETION_TYPE=@OLD_COMPLETION_TYPE*/; /*!50530 SET @@SESSION.PSEUDO_SLAVE_MODE=0*/; [root@ZST1 logs]#
语句操作过程：开启显式事务1，往repl_innodb表操作id=1(ins)、2(upd)、3(del)、4(ins)的记录-->开启事务2，往repl_myisam表写入id=1的记录，提交事务2-->开启事务3，更新repl_myisam表id=2的记录，提交事务3-->开启事务4，删除repl_myisam表id=3的记录，提交事务4-->开启事务5，往repl_myisam表写入id=4的记录，提交事务5-->提交显式事务1
binlog中我们可以看出，是按commit顺序，先记录repl_myisam的操作，然后记录repl_innodb的操作^_-
我们再来查看从库上的binlog
[root@ZST2 logs]# mysqlbinlog -v --base64-output=decode-rows mysql-bin.000003 /*!50530 SET @@SESSION.PSEUDO_SLAVE_MODE=1*/; /*!50003 SET @OLD_COMPLETION_TYPE=@@COMPLETION_TYPE,COMPLETION_TYPE=0*/; DELIMITER /*!*/; # at 4 #171201 22:54:07 server id 1333306 end_log_pos 123 CRC32 0xb5872a85 Start: binlog v 4, server v 5.7.19-log created 171201 22:54:07 at startup # Warning: this binlog is either in use or was not closed properly. ROLLBACK/*!*/; # at 123 #171201 22:54:08 server id 1333306 end_log_pos 154 CRC32 0x2f0d74be Previous-GTIDs # [empty] # at 154 #171201 23:01:19 server id 1323306 end_log_pos 219 CRC32 0x227de897 Anonymous_GTID last_committed=0 sequence_number=1 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 219 #171201 23:01:19 server id 1323306 end_log_pos 282 CRC32 0x9bd4c099 Query thread_id=13 exec_time=322 error_code=0 SET TIMESTAMP=1512140479/*!*/; SET @@session.pseudo_thread_id=13/*!*/; SET @@session.foreign_key_checks=1, @@session.sql_auto_is_null=0, @@session.unique_checks=1, @@session.autocommit=1/*!*/; SET @@session.sql_mode=524288/*!*/; SET @@session.auto_increment_increment=1, @@session.auto_increment_offset=1/*!*/; /*!\C utf8 *//*!*/; SET @@session.character_set_client=33,@@session.collation_connection=33,@@session.collation_server=33/*!*/; SET @@session.lc_time_names=0/*!*/; SET @@session.collation_database=DEFAULT/*!*/; ==================== repl_myisam表写入id=1的记录，重复key进行update操作，Start ==================== BEGIN /*!*/; # at 282 #171201 23:01:19 server id 1323306 end_log_pos 347 CRC32 0xe221b9d8 Table_map: `replcrash`.`repl_myisam` mapped to number 124 # at 347 #171201 23:01:19 server id 1323306 end_log_pos 425 CRC32 0x32dc6ddc Update_rows: table id 124 flags: STMT_END_F ### UPDATE `replcrash`.`repl_myisam` ### WHERE ### @1=1 ### @2='s1062-1' ### @3='s1062-1' ### SET ### @1=1 ### @2='m1062-1' ### @3='m1062-1' # at 425 #171201 23:01:19 server id 1323306 end_log_pos 489 CRC32 0x8e56c4de Query thread_id=13 exec_time=322 error_code=0 SET TIMESTAMP=1512140479/*!*/; SET @@session.sql_mode=1436549152/*!*/; COMMIT /*!*/; # at 489 ==================== repl_myisam表写入id=1的记录，重复key进行update操作，End ==================== ==================== repl_myisam表更新id=2、删除id=3的记录Start ==================== update/delete 在从库没找到记录，没有做任何处理，相当于skip ==================== repl_myisam表更新id=2、删除id=3的记录End ==================== #171201 23:01:52 server id 1323306 end_log_pos 554 CRC32 0x0401a98f Anonymous_GTID last_committed=1 sequence_number=2 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 554 #171201 23:01:52 server id 1323306 end_log_pos 617 CRC32 0xb8743fdf Query thread_id=13 exec_time=289 error_code=0 SET TIMESTAMP=1512140512/*!*/; SET @@session.sql_mode=524288/*!*/; ==================== repl_myisam表写入id=4的记录Start ==================== BEGIN /*!*/; # at 617 #171201 23:01:52 server id 1323306 end_log_pos 682 CRC32 0x8dc365d7 Table_map: `replcrash`.`repl_myisam` mapped to number 124 # at 682 #171201 23:01:52 server id 1323306 end_log_pos 738 CRC32 0x773e7fc0 Write_rows: table id 124 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_myisam` ### SET ### @1=4 ### @2='m1062-4' ### @3='m1062-4' # at 738 #171201 23:01:52 server id 1323306 end_log_pos 802 CRC32 0x221bf96b Query thread_id=13 exec_time=289 error_code=0 SET TIMESTAMP=1512140512/*!*/; SET @@session.sql_mode=1436549152/*!*/; COMMIT /*!*/; # at 802 ==================== repl_myisam表写入id=4的记录End ==================== #171201 23:00:32 server id 1323306 end_log_pos 867 CRC32 0x4e47e1fc Anonymous_GTID last_committed=2 sequence_number=3 rbr_only=yes /*!50718 SET TRANSACTION ISOLATION LEVEL READ COMMITTED*//*!*/; SET @@SESSION.GTID_NEXT= 'ANONYMOUS'/*!*/; # at 867 #171201 23:00:04 server id 1323306 end_log_pos 930 CRC32 0x081b8094 Query thread_id=13 exec_time=397 error_code=0 SET TIMESTAMP=1512140404/*!*/; SET @@session.sql_mode=524288/*!*/; ==================== repl_innodb表操作id=1、4的记录(2、3被skip)，在一个事务中Start ==================== BEGIN /*!*/; # at 930 #171201 23:00:04 server id 1323306 end_log_pos 995 CRC32 0x9797a0bd Table_map: `replcrash`.`repl_innodb` mapped to number 123 # at 995 #171201 23:00:04 server id 1323306 end_log_pos 1073 CRC32 0xe457ab20 Update_rows: table id 123 flags: STMT_END_F ### UPDATE `replcrash`.`repl_innodb` ### WHERE ### @1=1 ### @2='s1062-1' ### @3='s1062-1' ### SET ### @1=1 ### @2='m1062-1' ### @3='m1062-1' # at 1073 #171201 23:00:32 server id 1323306 end_log_pos 1138 CRC32 0x6fd579ab Table_map: `replcrash`.`repl_innodb` mapped to number 123 # at 1138 #171201 23:00:32 server id 1323306 end_log_pos 1194 CRC32 0x6c5678ef Write_rows: table id 123 flags: STMT_END_F ### INSERT INTO `replcrash`.`repl_innodb` ### SET ### @1=4 ### @2='m1062-4' ### @3='m1062-4' # at 1194 #171201 23:00:32 server id 1323306 end_log_pos 1225 CRC32 0x594d2812 Xid = 45 COMMIT/*!*/; ==================== repl_innodb表操作id=1、4的记录(2、3被skip)，在一个事务中End ==================== SET @@SESSION.GTID_NEXT= 'AUTOMATIC' /* added by mysqlbinlog */ /*!*/; DELIMITER ; # End of log file /*!50003 SET COMPLETION_TYPE=@OLD_COMPLETION_TYPE*/; /*!50530 SET @@SESSION.PSEUDO_SLAVE_MODE=0*/; [root@ZST2 logs]#
它也是按照主库的binlog的顺序，先提交repl_myisam操作，再提交repl_innodb操作。对于insert在从库有重复记录的，从库update为主库的值；对于update/delete在从库没找到记录的，不做任何处理，相当于skip~
When slave_exec_mode is IDEMPOTENT, a failure to apply changes from RBL(Row-Based Logging) because the original row cannot be found does not trigger an error or cause replication to fail. This means that it is possible that updates are not applied on the slave, so that the master and slave are no longer synchronized. Latency issues and use of nontransactional tables with RBR(Statement-Based Logging) when slave_exec_mode is IDEMPOTENT can cause the master and slave to diverge even further.
三、总结
正常情况下，应该是从库遇到错误就停止复制，然后人工去处理数据一致性问题。slave-skip-errors选项会导致SQL thread遇到配置中的错误后继续复制。滥用slave-skip-errors会在你全然不知的情况下导致主从数据不一致。
slave_exec_mode='IDEMPOTENT'，适合多主、循环、以及其他特殊复制场景。RBL(Row-Based Logging)环境下也会出现SBL(Statement-Based Logging)场景中主从数据不一致，复制状态正常
遇到复制中断第一时间要想怎么满足这个复制，而不是去跳过这个事务(･ω･)